© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 768
    9. 8-RES_FS768-JB
    11. 681
    Prev
    Next
    INDEX
    Explanations

    instances of change or alteration in a context

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    658
    +0.11
    0.5%
    189
    +0.11
    0.5%
    113
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.15
    642
    +0.11
    0.07
    140
    +0.10
    0.10
    Negative Logits
    amina
    -0.87
    ç«
    -0.66
    vern
    -0.64
    gins
    -0.64
    True
    -0.63
    APH
    -0.62
    acht
    -0.62
     Huck
    -0.61
    ¯¯¯¯
    -0.60
    ngth
    -0.60
    POSITIVE LOGITS
     drastically
    0.96
     tack
    0.87
     gears
    0.87
     radically
    0.85
     wording
    0.84
     dramatically
    0.82
     effected
    0.80
    iating
    0.78
    iations
    0.75
     perceptions
    0.75
    Activations Density 4.099%

    No Known Activations